Labour won a majority of seats in England alone in 2005 despite polling fewer votes than the Conservatives. Broadly speaking the electoral system tends to favour Labour when Labour does well and to have a more neutral effect otherwise.
Well that was more to do with Blair's 'Heineken' effect and was probably more of an exception to the rule, but broadly speaking I'd agree. It's worth noting that in 2010 when compared to 1992, as a % of seats won in England, both the Tories and Labour were generally at the same level (both lower, but similarly apart) and in 2010, the Tories did better in Wales than they did in 1992. Scotland is the exception of course and combined with the Lib Dems ensured that the election didn't completely mirror 1992.
Given that I think it's the most comparable election, particularly demographically then I would argue that because Blair wasn't there (rather than anything Cameron did), the association built up from 1994-2007 was 'broken'; most areas of the country reverted back to where it was before Blair. (1992 was itself a partial unwinding from Thatcher)
I think that's our starting point.
So if polls swing to produce a 2005 vote share scenario, I don't think that we will necessarily see a 2005 style spread of seats throughout the country.